PowerStore—The Power of Clustering
Mon, 27 Feb 2023 14:01:20 -0000
|Read Time: 0 minutes
Overview
PowerStore is designed to be a continuously modern storage platform. Its adaptable architecture allows for deploying appliances into a single- or multi-appliance cluster. Configuring a multi-appliance cluster with two to four appliances enables additional powerful functionality that you can be leveraging in your environment today!
PowerStore clustering capabilities are designed to simplify administration, provide integrated intelligence, and add flexibility, enabling multiple appliances to function as a single cohesive system. This blog discusses many of the benefits of deploying a multi-appliance cluster.
Simplicity
Configuring and managing a multi-appliance cluster is designed to be as simple as managing a single appliance. For both single- and multi-appliance clusters, storage administration is accomplished through a single user interface. The HTML5-based PowerStore Manager GUI provides an easy-to-use interface for management actions and monitoring operations that are crucial to an organization’s needs. There is no additional learning curve to manage your multi-appliance cluster.
Many PowerStore objects, such as protection policies, remote systems, vCenter and VASA provider connections, vVol storage containers, and hosts, exist at the cluster level. These objects can be used on all appliances, regardless of the size of the cluster. This alleviates the need to repeat operations on each appliance in the cluster.
For example, host registration for volume access needs to be completed just once. Hosts can be configured to have active initiators to one, some, or all appliances within the cluster, depending on your access requirements. This also enables volumes to be migrated between appliances within the cluster without any added management overhead. This can be used in situations such as if an appliance runs low on capacity, and when making service level changes, consolidating workloads, and more.
Most settings are also applied at the cluster level. These include PowerStoreOS code upgrades, security settings, network settings, user management, support materials, and more. Alerts, events, jobs, and audit logs are consolidated into a centralized area for all appliances in the cluster, eliminating the need to monitor each appliance individually. Performance and capacity metrics are available at the cluster, appliance, and resource levels. This provides the administrator with multiple levels of granularity needed for different tasks.
In multi-appliance configurations, each appliance has its own set of initiators for volume access. This level of granularity enables the highest level of control, providing the flexibility to connect specific hosts to specific appliances. However, if you want to connect a host to all appliances, you don’t need to go through the tedious effort of connecting to each appliance individually. PowerStore offers the option to configure a Global Storage Discovery IP address. This is a single, global, floating storage IP address that can be used to discover all paths for all the appliances in the cluster.
Intelligence
PowerStore includes integrated intelligence that is used to determine the initial placement of a new volume. This is known as the Resource Balancer, which is designed to streamline operations with an integrated machine-learning engine and seamless automation. Resource Balancer works at both the cluster and appliance levels. When a volume is provisioned, the default selection for placement is “Auto.” This setting allows the Resource Balancer to determine the best appliance for the new volume, depending on capacity metrics and configured host access. You can always maintain full control of volume placement by selecting a specific appliance for the volume as well.
The PowerStore active/active architecture means that volumes can always be accessed through both nodes of an appliance. When a volume is attached to a host, Resource Balancer also selects which node within the appliance to advertise as optimized for host access to that volume. This is known as node affinity. It enables both nodes within the appliance to actively service I/O simultaneously to multiple volumes, so all available hardware is efficiently utilized.
Workload characteristics might evolve over time and cause imbalances between nodes within an appliance. PowerStore features dynamic node affinity, which enables automatic balancing of node affinity of block storage resources between nodes. Dynamic node affinity allows the system to maintain relatively consistent utilization, latency, and performance between both nodes of an appliance. This intelligent architecture enables both seamless load-balancing and high availability.
PowerStore also allows for nondisruptive migration of volumes, volume groups, and vVols across the cluster. As capacity and performance characteristics and requirements change, users can initiate manual or assisted resource migrations of resources from one appliance to another. When resources are migrated, all associated storage objects, such as snapshots and thin clones, are also moved to the same destination.
The system periodically monitors storage resource utilization across all appliances within the cluster. As storage consumption increases over time, an appliance might start to run out of available capacity. In this scenario, the system generates migration recommendations based on factors such as drive wear, appliance capacity, and health. If the administrator accepts the recommendation, a migration session is automatically created. The PowerStore cluster can do all the planning for you!
Flexibility
PowerStore clusters offer flexibility by providing the ability to scale up, out, and down as needed. The initial cluster can be created with anywhere from one to a maximum of four appliances. If the maximum appliance count has not been reached, you can add appliances to the cluster at any time after the initial configuration without any disruption. The additional appliances can be used to add capacity, increase performance, and expand limits.
Within a cluster, you can mix appliances with different configurations such as models, drives, I/O modules, and fault-tolerance levels. NVMe expansion enclosures can be added to specific appliances within the cluster if additional capacity is needed. This allows each appliance in the cluster to have its own individual configuration that’s tailored for its specific use
Administrators can tell PowerStore to evacuate storage resources such as volumes, volume groups, and vVols from an appliance. This operation can be useful in situations where an appliance needs to be powered off for maintenance or removed from a cluster, or when migrating to a new appliance.
Appliances can just as easily be removed from a cluster. For example, after migrating data from one appliance to another, you might want to decommission or repurpose the original appliance. After ensuring that all the data is migrated, the appliance can be safely removed from the cluster. After the appliance is removed, it is reverted to factory settings so it’s ready to be configured as a new cluster, added to an existing cluster, or powered off.
Clustering can be used as an end-to-end life cycle management strategy to make operations such as hardware refreshes painless. The new appliance can be joined to the existing cluster without any impact, enabling both the old and new appliances to be used together. The existing data can be seamlessly and nondisruptively migrated from the old appliance to the new one. The migration can be done either incrementally over time or all at once. Once all the data is migrated, the old appliance can be repurposed or removed from the cluster. All these features and benefits of PowerStore clustering provide you with essential investment protection.
Conclusion
The PowerStore continuously modern storage architecture allows for deploying appliances into a single- or multi-appliance cluster with minimal complexity. PowerStore multi-appliance clusters provide many benefits and advantages, with simplified configuration and administration, integrated intelligence, and increased flexibility.
Resources
Author:
Wei Chen, Senior Principal Engineering Technologist
LinkedIn